Optimization Methods - Overview
Introduction
Overview of optimization problems in linear algebra: minima, maxima, saddle points, and the finite element method.
Key Concepts
Minima, Maxima, and Saddle Points
- Critical points: ∇f = 0
- Second derivative test
- Hessian matrix
- Constrained optimization
Quadratic Functions
- f(x) = ½xᵀAx - bᵀx
- Minimum when A is positive definite
- Connection to least squares
Lagrange Multipliers
- Constrained optimization
- KKT conditions
Finite Element Method
- Variational formulation
- Weak form of PDEs
- Stiffness matrix
- Applications to engineering
Applications
- Machine learning optimization
- Engineering design
- Structural analysis
- Computational physics
References
- Strang Chapter 6.1, 6.5
- Deep Dive